Add durable receipt auditing for guarded Antigravity backend dispatch - #1804
Closed
groupthinking with Copilot wants to merge 4 commits into
Closed
Add durable receipt auditing for guarded Antigravity backend dispatch#1804groupthinking with Copilot wants to merge 4 commits into
groupthinking with Copilot wants to merge 4 commits into
Conversation
Contributor
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Co-authored-by: groupthinking <154503486+groupthinking@users.noreply.github.com>
![vercel[bot]](https://avatars.githubusercontent.com/in/8329?s=60&v=4){kind=small}
Co-authored-by: groupthinking <154503486+groupthinking@users.noreply.github.com>
Copilot
AI
changed the title
[WIP] Add guarded Antigravity backend to Agent Factory
Add durable receipt auditing for guarded Antigravity backend dispatch
Sep 8, 2026
…e.total_tokens` in the `input_tokens` field, mislabeling token accounting (input = input+output).
This commit fixes the issue reported at src/youtube_extension/services/agents/adapters/agent_orchestrator.py:522
## Bug
In `execute_antigravity_backend` (`src/youtube_extension/services/agents/adapters/agent_orchestrator.py`, ~line 522), the durable audit-store append was populated as:
```python
total_tokens = receipt.usage.get("total_tokens")
output_tokens = receipt.usage.get("output_tokens")
self._audit_store.append(
...
input_tokens=(int(total_tokens) if isinstance(total_tokens, (int, float)) else None),
output_tokens=(int(output_tokens) if isinstance(output_tokens, (int, float)) else None),
)
```
The `PipelineAuditStore.append` signature (`src/youtube_extension/services/pipeline_audit_store.py`, lines 29-40) records `input_tokens` and `output_tokens` as separate prompt vs. completion counts. The canonical mapping elsewhere (`hybrid_processor_service.py` lines 321-336) derives `input_tokens` from the prompt token count and `output_tokens` from the completion token count.
Here `input_tokens` was fed `usage.get("total_tokens")`, which is the sum input + output. That mislabels the durable receipt: `input_tokens` ends up equal to input+output, corrupting cost/usage accounting derived from the audit trail.
## Trigger
Any completed Antigravity backend dispatch where the provider `usage` dict reports `total_tokens` (e.g. `{"input_tokens": N, "output_tokens": M, "total_tokens": N+M}`): the audit record writes `input_tokens = total_tokens` instead of the real prompt count.
## Fix
Read `usage.get("input_tokens")` (the actual prompt-token key, consistent with the sibling `output_tokens` key already read directly from `usage`) and record that into the audit store's `input_tokens` field:
```python
input_tokens = receipt.usage.get("input_tokens")
output_tokens = receipt.usage.get("output_tokens")
...
input_tokens=(int(input_tokens) if isinstance(input_tokens, (int, float)) else None),
```
`total_tokens` remains used for budget-limit enforcement in `antigravity_backend.py`, which is unaffected. The existing test only sets `usage: {"total_tokens": 321}` and does not assert on recorded token values, so it neither breaks nor catches this — a follow-up test asserting per-direction token recording would be worthwhile.
Co-authored-by: Vercel <vercel[bot]@users.noreply.github.com>
Co-authored-by: groupthinking <garveyht@gmail.com>
Contributor
|
Important Review skippedBot user detected. To trigger a single review, invoke the ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: ASSERTIVE Plan: Advanced Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Contributor
Dependency Review✅ No vulnerabilities or license issues or OpenSSF Scorecard issues found.Snapshot WarningsEnsure that dependencies are being submitted on PR branches and consider enabling retry-on-snapshot-warnings. See the documentation for more information and troubleshooting advice. Scanned FilesNone |
Contributor
🔍 PR Validation |
groupthinking
approved these changes
Sep 12, 2026
groupthinking
marked this pull request as ready for review
September 12, 2026 10:06
Contributor
|
Closing as unauthorized adjacent cut. Agent Factory / Antigravity is not the Origin G.A.T.E. cut. CONFLICTING vs rewritten |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Canonical issue
Linked by automation.
Outcome
Adds an optional, guarded Google Antigravity managed backend path in Agent Factory with durable receipt persistence at orchestration time. MCP/tool guardrails remain strict and audit records capture execution metadata without persisting task context.
Scope
AgentOrchestratornow accepts an optionalaudit_storeand records a durable entry onexecute_antigravity_backend.Risk
audit_storehook and associated test.Verification
List exact automated and manual checks, tied to the current head SHA.
Production evidence
Not applicable: backend orchestration/internal audit-path change only; no production surface or deployment behavior change.
Agent handoff